Anybody (Else) Use White Resin?

[C855B]

I'm having a little bit of an issue and need to compare notes. I seem to be experiencing what I call "bloom" - certain portions of parts render larger than the data and have somewhat rounded appearance, especially around supports. It is especially apparent on thin walls, where 0.010" will render to be 0.015" or even 0.020". I know the machine can do better, these windows print accurately, even to the square cross-sections of the very thin frames and muntins (0.010" & 0.015"):



What I am seeing seems to be similar to photographic overexposure. I currently have layers at 0.2mm and 6 seconds, the windows were 0.3 and 6, but I was having the "bloom" problem in the first run of the other bits at the same settings at the windows. Should I cut back on exposure?

Yeah, I can play with settings, it's only more time and very little resin, but another thought - I'm using white (Elegoo in this case). Is there possibly luminescence in white contributing to the overexposure? Gray and black (AnyCubic) won't arrive until next week to verify this postulate.

[Chris333]

For another 50 cents you might as well try it with lower exposure times.

[Stephane]

I had this same problem with anycubic grey, it just expanded much more than the green transparent.  I too just kept reducing exposure until supports started failing, and yet still had problems with blooming, especially around supports and facing the build plate (my long thread on this is "switched from anycubic green to gray.. what variable is wrong?")

Another possible solution, completely untested, might be to increase the distance between the build plate and your parts.  If too close to the build plate, just maybe the initial long exposures is cooking enough of the next layers?  Seems a long shot though.

Anyway, in my case, I just switched back to Anycubic Green transparent.  It gives me the best results, and I paint it just like I would the grey.

I am really curious though if you find a solution!!

[rodsup9000]

Mike,

 
 Are printing them vertical or horizontal?? I've found that I get better prints vertical.

 

[C855B]

Are you printing them vertical or horizontal?? I've found that I get better prints vertical.

Horizontal, 45°, and 60°. I'll try vertical next round and see if that helps. Batch in the printer right now at 15°, 0.03mm with 5 secs.

FWIW, the Kelso windows were printed at 45°.

[rodsup9000]

Another thing is to add your own support and don't use auto-support. I just do the bare minimum to get it to print.
When adding the supports, try to keep them on the faces that will be on the inside of the building and then you won't have to be so careful when removing the supports. When I do windows, I just peel them off the supports and some may need a little sanding on the bottom inside edge to fit into the window openings. 

[C855B]

Oh, no problem with the windows. They worked nearly immediately, the only bug on the first run was not enough exposure time for the rafts. It's the signal heads that are driving me crazy with the bloom. I'm working on separate parts that press-fit and the object growth is vexing. Latest try with 5 second exposure is curing, so I'll know more in the morning. In the meanwhile I'll create a .photon file with everything vertical and see how that fares when back in the studio tomorrow.

The auto-support issues were evident going in. I sometimes try it, maybe keep some, remove bad decisions and add in spots I anticipate as a problem. But more often than not I place supports myself.

[Lemosteam]

One of the early comments was lack of resin runoff all the way down to the cured surface.  @Chris333 likes to pull the parts off before curing (I get it about the thin cross section) but maybe there is excess resin still layering the print-cured part that gets cured atop the print when the UV station is used, making the parts rounded and "bloomed".

I might try an ultrasonic brush or cleaner with resin cleaner inside to remove all traces of excess resin from the part prior to the curing "oven".

This is one reason I am waiting for the higher res LCD, hoping that the LCD will control the exposure edges better.

[C855B]

One of the early comments was lack of resin runoff all the way down to the cured surface.  @Chris333 likes to pull the parts off before curing (I get it about the thin cross section) but maybe there is excess resin still layering the print-cured part that gets cured atop the print when the UV station is used, making the parts rounded and "bloomed". ...

Yes, worth considering. I am "pressure washing" the parts with alcohol from a laboratory wash bottle. Those four years of chem lab finally paid off.    I'll pay more attention to the areas around the supports on the next batch.

That said, cutting back on exposure time made a difference. Some of the 0.010 walls are now mic-ing out at ~0.012. Batch in the printer at the moment are vertical, per Rodney's @rodsup9000 advice. We'll see what happens there.

Here's a result of last night's test, hastily painted, and not loaded with the LEDs. I'm not entirely happy with the fitment of the back, and it's not as delicate as I would like, some of it due to the bloom. We'll continue to work on it. (Not taking credit for the relay box base - it's Showcase Miniatures.)



EDIT: Vertical-orientation prints just finished. Appearances are that this and 5-second exposure time fixed it. Too soft for calipers right now, so I'll know more after curing.

[C855B]

Confirmed - vertical orientation and 5 seconds for exposure yields a useful result. "Walls" are still +0.002" over the data, so I will adjust the art accordingly. I also slightly suspect the 100-second times for the raft layers might be pre-exposing anything at the bottom of the shortest supports. I'll cut that back to 80 seconds, the original 50 secs were detaching from the platform.

[Chris333]

I believe that Photon will always grow parts slightly. I've heard about a program that will shrink models to make up for this.

[peteski]

Confirmed - vertical orientation and 5 seconds for exposure yields a useful result. "Walls" are still +0.002" over the data, so I will adjust the art accordingly. I also slightly suspect the 100-second times for the raft layers might be pre-exposing anything at the bottom of the shortest supports. I'll cut that back to 80 seconds, the original 50 secs were detaching from the platform.

I don't  have first-hand knowledge of these printers, but I do know that they use an LCD screen.  Those have pixels, and printing is done in pixel-size increments. So, if the size of size of the item falls partially between the pixels, it will likely be rounded up to the next pixel increment.  Maybe that is a partial explanation of the "bloom"?

[narrowminded]

Confirmed - vertical orientation and 5 seconds for exposure yields a useful result. "Walls" are still +0.002" over the data, so I will adjust the art accordingly. I also slightly suspect the 100-second times for the raft layers might be pre-exposing anything at the bottom of the shortest supports. I'll cut that back to 80 seconds, the original 50 secs were detaching from the platform.

When using the extended cure times for the base supports, if you're using something like 8 layers, it will be eight at the layer height you've defined so is very unlikely that you've over cured the supports at the part attachment point as they are likely longer than that.  I have definitely experienced dimension growth in the X and Y planes and in the .002"/ .003" range that you've experienced.  Where it mattered I successfully adjusted the drawing to compensate.  The Z plane has generally been very good. 

[peteski]

The Z plane has generally been very good.

Likely because it is mechanically controlled but the stepper motor.

[Lemosteam]

I believe that Photon will always grow parts slightly. I've heard about a program that will shrink models to make up for this.

The issue with proportional scaling of the models, is that holes will become smaller as well and you will still have to modify the nominal design intent model.

I've said this so many times now.  I want a damn printer that can print to my design intent (within reasonable tolerances of course).  Unfortunately FUD/FXD is the only medium to do this (when the parts are smaller than 100mm  ) at this point from everything I have seen here.

This is why I will likely wait for the higher res LCD before I jump in.